Hi all, my latest blog post that some of you might find of interest - any thoughts welcome! Most teams can tell you their AI workflow passed evaluation. Fewer can tell you what "passed" actually required, or what happens when it fails. I built a six-dimension rubric that closes that gap: what's scored, what's gated, and why the difference matters more than the score itself. Full breakdown and the reasoning behind it in the article below. linkedin.com/posts/simonfreedman1_most-teams-can-tell-you-their-ai-workflow…?rcm=…&…

Love the article! Do you have any pointers on how to run these tests systematically? I'm having trouble figuring out the mechanics and experience that I otherwise get with tests for code. • Making them repeatable • Making them reproducible • On what cadence to run them • Managing costs for tests

Hey @Wasif Hyder, Thanks for the kind words! That's actually something I plan to cover in a future blog post so feel free to subscribe to the LinkedIn newsletter (free) and you'll be the first to hear when I do 😉